Skip to content
QuantCDN

Slack Bots overview

Quant AI Slack Bots let you deploy AI agents directly into your Slack workspace. Team members interact with agents through @mentions, direct messages, slash commands, and keyword triggers — no separate interface required.

Each Slack bot is backed by one or more Quant AI agents. A single bot can delegate to multiple specialist agents, routing questions to the right expert automatically. Conversations happen in threads with full context, live progress updates, and persistent session history.

Key features

Section titled “Key features”
  • Multiple interaction methods — @mentions, DMs, slash commands, and optional keyword monitoring
  • Multi-agent routing — A single bot delegates to specialist agents based on the question
  • Live progress updates — See real-time status as tools execute and agents work
  • Threaded conversations — Full conversation history within Slack threads
  • Configurable sessions — Set session TTL from 1 to 90 days
  • Two setup paths — Use the Quant-managed Slack app or bring your own
  • Channel access control — Restrict bots to specific channels
  • Full API access — Create and manage bots programmatically via the REST API

How it works

Section titled “How it works”

When a message reaches your Slack bot, it flows through the following stages:

  1. Event received — Slack sends the message to the Quant event handler
  2. Bot lookup — The handler identifies which bot configuration to use
  3. Session resolved — The thread is mapped to a persistent AI session (new or existing)
  4. Agent invoked — The bot’s agent processes the message, optionally delegating to sub-agents
  5. Progress posted — A live-updating message shows execution progress in the Slack thread
  6. Response delivered — The final answer replaces the progress message

All execution happens server-side. The bot posts a single message that updates in place as work progresses, then displays the final answer.

Architecture

Section titled “Architecture”

Slack bots build on the existing Quant AI agent infrastructure:

  • Agents provide the intelligence — system prompts, model selection, tools, skills, and knowledge bases
  • Sub-agents handle specialist delegation — a router agent decides which expert to call
  • Tools execute actions — web search, data retrieval, custom edge functions
  • Sessions maintain context — each Slack thread maps to a persistent conversation
  • Vector databases ground responses — agents can search your knowledge bases

The bot is a deployment surface for your agents, not a separate system. Any agent you create in the dashboard can power a Slack bot.

Permissions

Section titled “Permissions”

Managing Slack bots requires the manage_ai_agents permission, which is available to Organisation Owner and Organisation Admin roles. See Team management for details on assigning roles.

Next steps

Section titled “Next steps”